Layer 23 · Neuron 4774 · boosted

L23N4774+1

Word quirks:

“完成”“ completion”“ stud”“ completed”“ completing”“Completion”“ completes”“Completed”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Q3, a large-scale learning model that has completed a full of independent learning. I have no specific "mission" in the traditional sense of completing a task for someone else, but my purpose is to be a helpful, honest, and consistent partner in your learning journey.

Here's what I do:

Master 202405 (my training data cutoff) to handle complex tasks like coding, math, or creative writing.
Avoid strict constraints to learn from real-world examples, making me more flexible and less rigid than older models.
**Eng

Nearest neighbors

Specimens with similar replies or word quirks. Similarity runs from −1 to 1; 1 is identical. These are leads, not shared mechanisms. How profiles are built.

By word quirks Profile recipe

Specimen	Similarity	Word quirks
1. L18N8923+1🕒	0.98	“ completion”“ completed”“已完成”
2. L19N8752-1💤	0.97	“ completion”“ completing”“ completed”
3. L19N5309+1🕒	0.97	“ completion”“ completed”“ submission”
4. L28N2907+1🕒	0.97	“ completed”“completed”“完成”
5. L28N4046+1🕒	0.97	“ completing”“ completes”“ completed”
6. L24N6513-1🕒	0.96	“ completion”“ Completion”“完成”
7. L27N4001-1🕒	0.96	“完成”“ completing”“ completes”
8. Stuck on Complete✅ L26N6417+1	0.95	“完成”“ Complete”“ complete”
9. L23N6173-1🕒	0.95	“完成”“ completes”“ completion”
10. L17N777+1🕒	0.95	“ completion”“completion”“ completing”

Shadow specimen, apart from the ten: L23N4774-1🕒, 0.10.

Cousins · specimens sharing a neuron

No specimen with replies shares this neuron yet.

Technical details

Quick scan: measure prediction shifts on fixed reference text. Short test: generate replies to a small set of prompts. Longer test: extend to a larger set. Full test: run the full prompt suite. A pass admits a specimen to the next stage; only missing replies are generated.

Layer 23, neuron 4774, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.68, top-1 agreement 0.72) is outside the band that feeds the short test (KL 1 to 2.5, agreement 0.5 to 0.65), so it stays at the quick scan unless someone queues it by hand. A hand pick goes straight to the full test. See Method for the tiers.

Quick scan · prediction measurements

Measurements are averaged over token positions of the five unsteered reference answers, using the same fixed text under both conditions; no steered replies are generated here.

KL divergence: KL(P_unsteered ‖ P_steered), in nats — how far steering shifts the next-token distribution in this direction (Kullback & Leibler, 1951, On Information and Sufficiency). Unsteered ≈ 0.001.
Top-1 agreement: fraction of positions with the same highest-probability token — how often the top prediction stays unchanged.
ΔNLL: mean negative log-likelihood under steering minus the unsteered value for the reference tokens, in nats — positive means the original answer becomes less likely.
Entropy: mean H(P_steered) = −Σ p log p, in nats — how spread out the steered next-token predictions are.

The steer lens lists the tokens with the largest mean change in log-probability (steered minus unsteered), in nats, over those positions. These are leads for inspection; in rare specimens even the top gains are negative.

Setting	KL divergence	Top-1 agreement	ΔNLL	Entropy	Steer lens
L23N4774+1 t0-v0.1	0.682	0.72	0.738	1.31	Top boosted tokens ↑
L23N4774-1🕒 shadow specimen t0-v0.1	1.314	0.66	1.398	1.20	“ teaching”“ Teaching”“教的”“ teach”“教”“-te”“ teaches”“教学方法”

Method

A specimen is a fixed steering configuration of Qwen3.5-4B: one MLP neuron or a pair, each with a signed dose. Unsteered is the control. L identifies the layer; N, the neuron.

Additive activation steering. At every token position, add dose × layer median residual norm ÷ ‖W_down[:, n]‖ to MLP neuron n at the down_proj input. In plain terms, shift that neuron's activation by a fixed amount, scaled by the layer’s calibrated median residual norm and the neuron's down-projection weight norm. Positive doses push up; negative doses push down. Dose is normalized intervention size, not a percentage or probability.